Re: [PATCH 1/3] ext4: Fix quota accounting error with fallocate

[Mingming <cmm@us.ibm.com>]

On Thu, 2010-01-07 at 00:52 +0530, Aneesh Kumar K.V wrote:
> When we fallocate a region of the file which we had recently written,
> and which is still in the page cache marked as delayed allocated blocks
> we need to make sure we don't do the quota update on writepage path.
> This is because the needed quota updated would have already be done
> by fallocate.
> 
> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> ---
>  fs/ext4/ext4.h    |    2 ++
>  fs/ext4/extents.c |   21 +++++++++++++++++++++
>  fs/ext4/inode.c   |   35 +++++++++++++++++++++++------------
>  3 files changed, 46 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index af7b626..b98de17 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -1443,6 +1443,8 @@ extern int ext4_block_truncate_page(handle_t *handle,
>  extern int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
>  extern qsize_t *ext4_get_reserved_space(struct inode *inode);
>  extern int flush_aio_dio_completed_IO(struct inode *inode);
> +extern void ext4_da_update_reserve_space(struct inode *inode,
> +					int used, int quota_claim);
>  /* ioctl.c */
>  extern long ext4_ioctl(struct file *, unsigned int, unsigned long);
>  extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long);
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index 7d7b74e..3b6ff72 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -3132,7 +3132,19 @@ out:
>  		unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
>  					newblock + max_blocks,
>  					allocated - max_blocks);
> +		allocated = max_blocks;
>  	}
> +
> +	/*
> +	 * If we have done fallocate with the offset that is already
> +	 * delayed allocated, we would have block reservation
> +	 * and quota reservation done in the delayed write path.
> +	 * But fallocate would have already updated quota and block
> +	 * count for this offset. So cancel these reservation
> +	 */
> +	if (flags & EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE)
> +		ext4_da_update_reserve_space(inode, allocated, 0);
> +

Looks right to me, we are only updating the reserve space in the delayed
buffered IO case, but avoid updating the quota.

>  map_out:
>  	set_buffer_mapped(bh_result);
>  out1:
> @@ -3368,9 +3380,18 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
>  	/* previous routine could use block we allocated */
>  	newblock = ext_pblock(&newex);
>  	allocated = ext4_ext_get_actual_len(&newex);
> +	if (allocated > max_blocks)
> +		allocated = max_blocks;
>  	set_buffer_new(bh_result);
> 
>  	/*
> +	 * Update reserved blocks/metadata blocks after successful
> +	 * block allocation which had been deferred till now.
> +	 */
> +	if (flags & EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE)
> +		ext4_da_update_reserve_space(inode, allocated, 1);
> +
> +	/*
>  	 * Cache the extent and update transaction to commit on fdatasync only
>  	 * when it is _not_ an uninitialized extent.
>  	 */
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index c818972..3d1a1d6 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -1053,7 +1053,8 @@ static int ext4_calc_metadata_amount(struct inode *inode, sector_t lblock)
>   * Called with i_data_sem down, which is important since we can call
>   * ext4_discard_preallocations() from here.
>   */
> -static void ext4_da_update_reserve_space(struct inode *inode, int used)
> +void ext4_da_update_reserve_space(struct inode *inode,
> +					int used, int quota_claim)
>  {
>  	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
>  	struct ext4_inode_info *ei = EXT4_I(inode);
> @@ -1090,9 +1091,17 @@ static void ext4_da_update_reserve_space(struct inode *inode, int used)
>  	spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
> 
>  	/* Update quota subsystem */
> -	vfs_dq_claim_block(inode, used);
> -	if (mdb_free)
> -		vfs_dq_release_reservation_block(inode, mdb_free);
> +	if (quota_claim) {
> +		vfs_dq_claim_block(inode, used);
> +		if (mdb_free)
> +			vfs_dq_release_reservation_block(inode, mdb_free);
> +	} else {
> +		/*
> +		 * This is a request to cancel the reservation. So just
> +		 * update the resevation and cancel the quota reservation
> +		 */
> +		vfs_dq_release_reservation_block(inode, mdb_free + used);
> +	}
> 
>  	/*
>  	 * If we have done all the pending block allocations and if
> @@ -1292,18 +1301,20 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
>  			 */
>  			EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
>  		}
> -	}
> 
> +		/*
> +		 * Update reserved blocks/metadata blocks after successful
> +		 * block allocation which had been deferred till now. We don't
> +		 * support fallocate for non extent files. So we can update
> +		 * reserve space here.
> +		 */
> +		if ((retval > 0) &&
> +			(flags & EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE))
> +			ext4_da_update_reserve_space(inode, retval, 1);
> +	}
>  	if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE)
>  		EXT4_I(inode)->i_delalloc_reserved_flag = 0;
> 
> -	/*
> -	 * Update reserved blocks/metadata blocks after successful
> -	 * block allocation which had been deferred till now.
> -	 */
> -	if ((retval > 0) && (flags & EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE))
> -		ext4_da_update_reserve_space(inode, retval);
> -
>  	up_write((&EXT4_I(inode)->i_data_sem));
>  	if (retval > 0 && buffer_mapped(bh)) {
>  		int ret = check_block_validity(inode, "file system "

Looks good,
Reviewed-by: Mingming Cao<cmm@us.ibm.com>